[Profile picture of Ruben Verborgh]

Ruben Verborgh

Enabling Dataset Trustworthiness by Exposing the Provenance of Mapping Quality Assessment and Refinement

by Tom De Nies, Anastasia Dimou, Ruben Verborgh, Erik Mannens, and Rik Van de Walle

Assessing the trustworthiness of a dataset is of crucial importance on the Web of Data and depends on different factors. In the case of Linked Data derived from (semi-)structured data, the trustworthiness of a dataset can be assessed partly through their mappings. The accuracy with which schema(s) are combined and applied to semantically annotate data – as described by its custom mapping definitions – plays a determinant role to the dataset’s overall potential. The inherent value of mapping definitions is often neglected, as they are considered part of the implementation executing them. However, an approach was proposed to assess and refine such mapping definitions, which was proven to be more effective than assessing and refining the quality of a dataset directly. In this paper, we derive important metadata from mappings quality assessment and refinement in the form of provenance information. The provenance of these mappings enables us to assess the relative trustworthiness of the datasets they generate.

BibTeX Mendeley

Published in 2015 in Proceedings of the 4th International Workshop on Methods for Establishing Trust of (Open) Data.

Keywords: Linked Data, Web, metadata, provenance

Read this paper online

Cite this paper in your publications

Discuss this paper